abc_crypto

[Raghav Dixit @_raghavdixit_]

https://t.co/RVnYCtg2wO

[abc_crypto @abccrypto5]

@PeterLBrandt A timeless classic! My trading vastly improved once I started applying insights from this book. Once you played the marble game and get to that that aha moment you are on your way to becoming a better trader.

[abc_crypto @abccrypto5]

@cupcake_trader @Tony_BATtista Thanks for putting this together! In my experience paper trading takes you only so far. Something that looks good on paper usually doesn't survive in the real world.

[abc_crypto @abccrypto5]

@OldSaltyGrouch @Tony_BATtista @wearelossdog F yeah!

[abc_crypto @abccrypto5]

@stephen_yo @NFTfi I'm proud to have been part of this journey. I was a lender on @NFTfi doing over 500 loans lending 400+ ETH over this period. I am sad to see this chapter come to an end but that's life I guess. Not everything works out. It's been a fun journey I wish the NFTfi team the best.

[abc_crypto @abccrypto5]

@Tony_BATtista @15minofPham Looks like it's time to get the old gang back together and start a new firm! I'll move my TT account in a second if you guys do that!

[abc_crypto @abccrypto5]

@PrestonPysh I learned a lot about Bitcoin from you when I first got in the space years ago! Thank you for your service!

[abc_crypto @abccrypto5]

@mentatminds Test $150 first then zoom!

[abc_crypto @abccrypto5]

@CadeONeill A huge fan of Gavin Wood and invested in so many DOT parachains I’ve lost count. So when I saw TAO I was hoping that maybe this one might be the one?!

[abc_crypto @abccrypto5]

@const_reborn Right on, brother! This is nothing but a slight bump in the journey. What doesn't kill you makes you stronger!

[const @const_reborn]

https://t.co/IdB9DPr5u0

[abc_crypto @abccrypto5]

@bittensormax Ohh this would be so cool!

[Jose Caldera @josercaldera]

When we built Yanez Compliance, we were laser-focused on financial crime — sanctions screening, model validation, compliance automation. Hard, unglamorous, important work.

And it worked. We helped teams do in hours what used to take thousands of hours. We cut false positives. We made compliance programs defensible.

But while we were doing that, deepfake identities started to wreck havoc. LLMs exacerbated how easy it was to bypass identity verification. And it so happens we have been working on the technology already to help organizations prepare better for it.

The very core of identity verification as a tool for fraud prevention needs to evolve: deepfakes are turning identities into a vulnerability.

What we need are new primitives for trust. Proof of humanhood, and proof of uniqueness, paired with proper non repudiation of activity are the key.

So we put our team's combined expertise of 100 years in building: digital identities, biometrics, cryptography, blockchain decentralization to work.

Today we're rebranding. Our foundation is the same. Our ambition is bigger.

One Human. One Unique Human.

👉 https://t.co/Qpd22HI5Zu

$TAO @yanez__ai

[Lamida @LamidaGlobal]

https://t.co/GcREhSCUzu

[Yanez.ai @yanez__ai]

https://t.co/okutrfdxk2

[Jose Caldera @josercaldera]

Don’t ever shy away from a conversation, you may miss someone truly unique and remarkable. At Breakout I got to meet a living legend: John Fucking Caldwell. That is the guy with the hat in the picture below. We met on Monday evening at Breakout, and today we got to Go to his winery and had an amazing experience: 4 hours of wine and Bittensor. We spend so much time between agents and AI that we forget that what makes us humans and unique are the people we meet and we share our experiences with. Go Humans!

$Tao

[abc_crypto @abccrypto5]

@RBS_HODL That’s 120% growth year over year!

[seth bloomberg @bloomberg_seth]

There have been a lot of good replies to this post, but I started writing this yesterday so I decided to finish and post it anyway. But before getting into the core of the piece, I want to correct some inaccuracies and statements presented as facts that are incorrect. For example:

> “Daily emissions of 3,600 TAO are split programmatically: 41% to miners, 41% to validators, 18% to subnet owners.”
Subnet participants (owners, validators, miners) don’t directly receive TAO, they receive subnet tokens.

> “Chutes commands 14.4% of network emissions, the highest of any subnet.”
Right now, Chutes is getting 0% of network emissions. But, that doesn’t even really matter for them. They have a market cap of $126M with a liquidity pool that has $135M worth of liquidity. Network emissions primarily just add liquidity for subnet tokens and Chutes is extremely liquid. Miners still get rewarded for their work (since they’re rewarded with the subnet token) even when a subnet gets 0% network (TAO) emissions.

> “[Chutes has an] estimated external revenue of approximately $1.3-2.4 million per year.”
This data is out of date. Chutes earned $1.3M over the last 90 days, annualized that’s ~$5.2M.

> “When emissions halve again (projected late 2026 or 2027), either pricing roughly doubles, miners leave the subnet, or the gap between subsidy and revenue widens further.”
The next TAO halving is expected to be in December 2029. And again, network emissions (i.e., TAO emissions) do not directly go to miners. So the TAO halvings have no direct impact on miners.

> “A single team controls nearly a quarter of the network’s incentive distribution.”
You’re implying here that Rayon labs “controls” a quarter of the network’s incentive distribution. But, the network's incentive distribution is purely market-driven. Holders of TAO get to decide where TAO emissions are directed, they vote via the subnet tokens they hold.

As I mentioned earlier, there have been some good responses to specific parts of your piece, and some of your critiques hold true today. So, I’ll stay at a bit of a higher level here. Bittensor has a lot of moving parts to it, I’ll try to just run through how I see them all connecting.

Subnets are startups launched on Bittensor. For Bittensor to succeed, it’s not necessary that all or even the majority of subnets do well. Most will actually fail since they are startups. It only needs a handful of breakout winners to (1) prove the network model works, (2) generate real revenue, and (3) attract new crops of builders. We’ve seen the probability of each of these points becoming true increase over the past few months.

Templar is the latest example to show the network's core tech (i.e., incentive mechanisms) works. They were able to attract enough compute from global participants to train a 72B model over the internet. A year or so ago people basically thought this was impossible. Turns out if you create the right economic/tech mechanisms to fund interesting research you can actually get breakthroughs using distributed and decentralized methods. IOTA is in the same category here, just coming at decentralized training from a different angle and doing novel research in their own right.

Chutes has built a real, live product and is generating multiple millions of dollars in revenue. They’re at ~$5.2M in annualized revenue, and becoming more efficient as a business (they just put out a lengthy blog post describing their focus on efficiency). Whether you’re using VC funding to subsidize your business or tokens, eventually you have to become profitable/remove the subsidy. Chutes team is dialed in on this. It’s also worth noting here that Chutes is commoditizing access to inference/models. I don’t think it’s fair to say it’s trivial to run models locally at this point, you can still create a strong/defensible business in this domain. Based on the pockets of success we’ve seen so far on the network, we believe a handful of subnets can hit an average ARR of $30M by the end of this year. With these valuations and associated multiples, subnets can achieve valuations of ~$900M. Again, we don’t need every subnet to have this level of success, only a few need to break out and multiple already have motion.

As an aside (but thought I would include since you specifically discussed Chutes), all of the above notes on Chutes are focused on the demand side of the platform, but they also have a structural advantage on their supply side. Chutes provisions compute differently than every other inference provider in the market. Chutes has a permissionless and open supply side, they don’t have the operating costs of compute contract negotiation or sourcing/vetting compute providers. They have an onchain, continuous compute budget that is available for any compute provider that plugs into their network and performs well. Because Chutes leverages interruptable compute nodes, they will attract the lowest-cost compute, which will likely be a datacenter with idle hardware. These entities will take anything >$0, given they need to generate ROI on their hardware in a shrinking window of time. This lowers Chutes' cost of compute provisioning, giving them a structural advantage over their competitors. Traditional inference platforms have to commit capital to long-term compute contracts and take on that price and counterparty risk, or opt for shorter-term contracts that come with a premium (much more than what Chutes pays for the equivalent unit of compute).

It only takes a handful of subnets to just chew glass for the overall betterment of the ecosystem. Every single crypto platform goes through this same cycle of attracting early talent that helps smoothen out rough edges of building in the ecosystem, paving the way for new builders to more easily join and find success. There’s already been a class of Bittensor builders who have done just that. And what they build for the entire ecosystem will attract new builders just like every other crypto platform has shown to be true. I know of a couple subnets launching in the coming quarter that will be raising the bar on subnet quality and they all will benefit from other builders in the ecosystem that came before them.

And to that last point, every existing and future builder on Bittensor benefits from any singular success on the network. Subnets actually have a structural advantage that basically every other platform/ecosystem lacks. Because every subnet token is priced in TAO terms, every subnet benefits as the value of TAO increases. Take the recent run up in TAO price that was largely tied to the widely publicized achievements of Templar, a single subnet. Every single subnet on Bittensor saw their miner budget increase by ~70% over the past 30 days because of the success and public recognition of a single subnet. You don’t have this advantage if you aren’t built on Bittensor.

There’s actually a local flywheel that every subnet is trying to get started, which in aggregate, creates a really strong global network effect. For each subnet, all the participants are incentivized to make the subnet token more valuable: token holders (including subnet owners, miners, and validators) benefit from price appreciation, and higher token prices mean subnet owners have a bigger budget to pay miners with. The only way to make the subnet token go up is by acquiring TAO and staking it (i.e., swapping it) to the subnet and holding the token. The simplest design for this is token buybacks by subnet owners, but there will be many other demand drivers developed for subnet tokens this year. Every subnet is effectively racing and incentivized to acquire more TAO into its subnet liquidity pool. And every marginal subnet benefits as the value of TAO appreciates. It’s a novel design that ties individual application success to platform success.

At the end of that day, if you come to Bittensor and build something that genuinely gets people excited, whether it’s a consumer product or deep research, you can be successful. I believe it actually takes a mix of both “day 1” revenue generating subnets coupled with deeply technical subnets that attract bright minds that want to be a part of frontier research. Both are finding success right now (Templar, IOTA, Chutes and Targon are the perfect examples of each) and will continue to do so in the future.

[@jason @Jason]

$tao > $btc

[abc_crypto @abccrypto5]

@macrocrux Shit, no wonder payouts went down